Project 1 will follow up on initial observations that the diffuse character of these tumors is not simply due to the tumor cells migrating into the normal brain structures as had been previously thought but that the tumor actively recruits surrounding cells from the brain (brain stroma) and other sources into the tumor mass and induces their aberrant proliferation. These recruited stem/progenitor cells may function normally in the response to injury in a self-limiting manner. We will use lineage tracing combined with oncogenesis to genetically distinguish the progeny of the original tumor cells (marked by GFP expression) from the recruited cells. We will additionally use bioluminescence imaging to monitor the signaling activity of pathways known to drive stem/progenitor character in these recruited cells in slice preparations of tumors in situ. We have created reporter mice that express luciferase proportional to SHH, Wnt, and Notch signaling activity. In Aim 1 we will determine the location from which the recruited cells arise, specifically the bone marrow and brain contribution. Our preliminary data indicates that a high percent of cells within gliomas are not derived from the original cell of origin and are recruited from other locations. In Aim 2 we will determine whether loss of the known glioma tumor suppressors (INK4a, Arf, PTEN) will promote the incorporation and contribution of recruited cells to the tumor proper. In theory the recruited cells could acquire mutations independent from those found in the progeny of the cell of origin, and the cells making up the most malignant tumors could be derived from recruited cells. In Aim 3 we will determine if the Gli-luc mouse line can be used for bioluminescence imaging to visualize recruitment and SHH signaling in gliomas overtime. The signaling pathways that attract recruited progenitor cells to the tumor and maintain them in proliferative and undifferentiated state are unknown. Aim 4 we will determine if the SHH signaling and cell recruitment is a response to injury. Our preliminary data indicates that SHH signaling in gliomas is an in vivo phenomenon and that it may be a dysregulated response to injury that normally is tightly controlled temporally. We will systematically determine if SHH activation is part of a normal response to injury and if the alterations found in human gliomas that are known to induce gliomas in mice alter this effect. Project 2 in this proposal seeks to identify the genes and functions that enable metastatic colonization of the brain microenvironment. Brain metastasis indicates poor prognosis in several major malignancies. In breast cancer, brain metastasis is on the rise as progress is being made in controlling relapse to other sites. Brain metastasis requires that tumor cells break through a unique physiological barrier, the blood brain barrier (BBB). Once this is achieved, tumor cells must then interact with the brain parenchyma environment in order to colonize it. Despite these distinctive features, the underlying molecular and cellular mechanisms remain a mystery. Using methodology that we recently developed to identify organ-specific metastasis genes and functions, we will isolate brain-specific metastatic cell populations. (Aim1). Our source of malignant cells is pleura! fluids of breast cancer patients undergoing treatment at MSKCC. In-vivo selected brain-metastatic cell populations will be subjected to comparative transcriptomic analysis in order to identify genes that mediate brain metastatic colonization (Aim 2). Our preliminary results demonstrate the feasibility of this approach and have already provided a provisional brain-metastasis gene-expression signature shared by different patients. This gene list contains candidate mediators of BBB breakdown and tumor cell interaction with brain parenchyma and microglia for invasion, colonization and inflammation. We will functionally validate these genes as mediators of brain metastasis by means of genetic silencing and pharmacological inhibition. In parallel, we will seek clinical validation by analysis of clinical brain metastasis tissue and malignant cells from cerebrospinal fluid (Aim 2). Based on hypotheses raised by this cohort of candidate genes, we will investigate specific cellular and molecular events involved in the breakdown of the BBB during brain metastasis in an animal model (Aim 3). Also based on the cohort of brain metastasis genes, we will investigate tumor-microenvironment interactions mediating colonization of the brain parenchyma (Aim 4). To these ends, we will use several complementary approaches including newly developed in-vivo and ex-vivo functional assays and imaging techniques. This work will be pursued in close cooperation with Project 1 (Aims 2 and 4), and Project 3 (Aims 3 and 4), and the support of a multidisciplinary group of clinical collaborators and mentored technologies. Reciprocal interactions between cancer cells and their extracellular surroundings define the tumor microenvironment. By comparison to other tissues, the brain microenvironment represents a number of unique features, including a distinct extracellular matrix and several specialized host cells such as astrocytes, neurons, oligodendrocytes and the microglia. Moreover, the brain lacks a lymphatic system, and is considered an immunologically privileged organ, relying on the blood-brain barrier (BBS) to exclude toxic substances. However, the brain is an organ site that has been largely neglected in the field of tumor-host interactions, and thus there is a fundamental gap in our understanding of how the brain microenvironment contributes to tumorigenesis. The long-term objective of Project 3 will be to establish not only how tumors modify host cells in the brain for their own ends, but also how the normal brain environment responds to the initiation and progression of both primary tumors and metastases. The central hypothesis is that stromal cells, particularly macrophages and endothelial cells, provide key factors, such as matrix-degrading enzymes, which are co-opted by the tumor to promote its own growth and invasion. Guided by strong preliminary data, we will test this hypothesis by pursuing four specific aims. In Aim 1 we will determine the role of the matrix-degrading enzymes, heparanase and cathepsins, in the tumor microenvironment during gliomagenesis. Our preliminary results show local up-regulation of these enzymes by stromal cells in areas of glioma angiogenesis and invasion. We will test whether they are important in these two processes, using validated pharmacological and organotypic co-culture strategies. Our preliminary data also implicates these matrix-degrading enzymes in metastasis, and in Aim 2 we will elucidate the mechanism by which they contribute to metastatic homing, breaching of the blood-brain barrier, and survival of metastatic lesions. In Aim 3 we will establish whether macrophages in the brain microenvironment are tumor-promoting, by their selective ablation during gliomagenesis and metastasis. In Aim 4 we will further investigate the role of macrophages in the tumor microenvironment by identifying the gene expression signatures associated with conversion of a normal macrophage to a tumor-associated macrophage. To achieve these aims, we will use a comprehensive experimental strategy combining sophisticated mouse models, imaging, organotypic coculture, flow cytometry, microarray profiling, pharmacological approaches and analysis of patient samples. The research in this proposal will be accomplished through the close collaboration already established with Project 1 (Holland) and Project 2 (Massague). These experiments will provide vital insights into how the brain microenvironment influences tumor initiation, progression and metastasis, enabling us to devise and test future therapeutic strategies, that will have immediate relevance to patient health.